Technology

➚ MOVE FAST: Going big. The hype around generative AI is translating into eye-popping valuations. Chatbot startup Character.AI is in talks to raise a new round of funding that would value the company at $5 billion, up from just $1 billion earlier this year, according to Bloomberg.

➘ BREAK THINGS: Being too big. The Google antitrust trial will get spicy on Monday when Microsoft CEO Satya Nadella is expected to testify as a witness for the U.S. Justice Department. More than 20 years ago, it was the Windows maker in the crosshairs. French authorities, meanwhile, raided Nvidia’s local offices as part of a probe into alleged anti-competitive practices.

The CEO of Medium has a message for AI companies: Don’t scrape our content. Tony Stubblebine, who took over last summer, wrote in a Medium post what is probably the quote of the day: “Why would we let AI companies leach off our stories, just to turn around and spam our favorite corners of the Internet?”

It’s difficult to argue against that logic. But that world view is also pretty narrow. In addition to creating spam, large language models are potentially incredibly useful. Some people argue the internet itself has harmed our culture and institutions. Yet, it’s hard to imagine a world without it.

We might have to come to terms with the fact that our “favorite corners of the internet” aren’t permanent. And then, the question is what comes next?

Yusuf Mehdi is Microsoft’s chief consumer marketing officer and recently took over some of the duties of the company’s former lead product executive.

Q: GPT-5 will be more multimodal with audio and video. What’s your vision for how multimodal will progress?

A: We’re doing a lot of it already today. I had this scenario with my wife the other day. She sent me a picture of a piece of art. She was like, ‘hey, maybe we’ll buy it for birthdays or something. You remember it, right?’ And I wasn’t sure. So I took the picture and loaded it into Bing Chat. It came back and told me, this is the artist, this is the title, here’s a bit about the story behind this piece.

So then I was able to go back and say, ‘yeah, I remember that.’ That was a cool multimodal. It’s extra compute power. So it costs more money to serve those GPT queries. And then there’s a lot of work to ground it with the search engine, so that we look at the image and the text. You’ll see more and more people using it to walk around the world and learn things.

Q: How do you forecast the capabilities of these models so you can plan your products down the line?

A: We have a tight relationship with OpenAI. And the model isn’t the only thing. There’s actually a lot of things on top. At our February event, we announced our Prometheus model, which is our proprietary way of working with an LLM. You get better results if you get better prompts. So when someone does a search, rather than just putting that in there, we run AI on it, we compare it against the big index, because we do a lot of query disambiguation to say, ‘what’s a smart way to put that into the LLM?’ The LLM runs it, it comes back and says, ‘I think this is the answer.’ We check that against the web. And we’re like, ‘no, this is probably not going to be a great answer.’

So we’re doing at multiple levels of the technology stack, grounding, training, even before you do anything special in the model. We can spend quite a long time innovating, writing code, and shipping things to make it better even before we get to one of the next models.

Q: There is kind of an art to prompting. How does that change consumer behavior and the language that we use?

A: We’re having to retrain our brains because we were taught to dumb down our searches. Fewer keywords give you better search results. The average today is 2.6 keywords, that’s the average in 10, 20 billion queries every single day. In AI, you’re so much better off saying, ‘Give me a picture of a sunset at the edge of the equator, in autumn, where the moonlight is reflecting off and the sun is halfway down. And I want to see lots of blues and oranges.’ You can’t put any of that in the search engine. But that makes your AI prompt better. So we’re having to retrain our brain to feel comfortable asking for more specific things.

Q: I wonder if the younger generation will be better at that?

A: My daughter is a budding artist. She had to write her artist statement for her class at school. She asked, ‘Dad, do you think I can use an AI?’ She’s of that generation that is very good with technology. But she’s an artist. She’s not a techie. She came back and said, ‘I got something great. But it took me a long time to get the prompts right.’ She couldn’t believe how much work she had to do, telling it what she’s focusing on, what her passion is, what she wanted it to do. It forced her to think before writing. What am I really about? Because she’s more of an artist, she’s not great at writing these long docs about herself. But she was great about putting her ideas down.

There’ll be a class of people, maybe it’ll be a younger generation, who will understand these prompts. What points do I want to make? That’s why it’s not surprising that the people who are really using them today are writers, artists, coders, because it’s people who know how to create.

Increase in the number of likes and shares that posts from Russian, Chinese, and Iranian propaganda outlets received on X (formerly Twitter) during the first 90 days after Elon Musk removed “state-affiliated” labels from their accounts, according to a new analysis by NewsGuard. One of the outcomes of Musk’s Twitter takeover is that the platform is now friendlier to foreign governments.

The academic journal Nature surveyed 1,600 researchers about the effect that artificial intelligence will have on science. About half of the respondents study or develop AI themselves, and they were predictably bullish about its potential upsides, especially how it could boost productivity. But they were also forthcoming about the downsides, like how AI may increase the amount of junk science that gets published.

Chinese merchants selling counterfeit designer goods are hosting full-blown fashion shows on TikTok. NBC News reporter Ben Goggin first spotted the phenomenon last weekend, and I began seeing the livestreams crop up on my own For You page over the last few days. During one broadcast, two thin women wearing black face masks modeled “Balenciaga” t-shirts they tossed on over black sports bras.

Producing counterfeits, or “replicas” as fans sometimes call them, has become a sophisticated business over the last few years, as manufacturers and merchants steadily improved their tactics. While once largely the purview of cheap sellers on New York City’s Canal Street, high-quality fakes can now sell online for $200 or more and are often made from the same materials as the real deal. They typically arrive in similar luxurious packaging, too.

Counterfeits of the most coveted items — a popular Chanel bag currently being sold in stores, for instance — might only be available by contacting a private seller on WhatsApp or Telegram. Other products, however, are easily found on Chinese platforms like AliExpress, and especially, DHGate. When I looked up the latter site on Google on Thursday, one of the first results was for a $34 knockoff of the viral Prada Cleo bag (the authentic version costs almost 100 times as much).

— Louise

• Democrats say it’s time for Menendez to go. Republicans say not so fast.
• The world’s biggest climate fund wants to get bigger. But even more is needed to help developing economies decarbonize.
• Fox slashed ad prices for the second Republican debate. One ad buyer told Semafor the race “has become a snoozer.”